Title :
Constructing scalable TTS system based on Corpus approach
Author :
Wei Zhang ; Zheng-hua Ling ; Li-Rong Dai
Author_Institution :
Dept. of Comput. Sci., Ocean Univ. of China, Qingdao
Abstract :
Pruning redundant synthesis instances or tailoring TTS voice font is an important issue of Corpus-based TTS. But pruning redundant synthesis instances, usually results in loss of non-uniform. In order to solve this problem, this paper proposes the concept of virtual non-uniform. According to this concept and the synthesis frequency of each instance, the algorithm named StaRp-VPA is constructed as to make up the loss of non-uniform. In experiments, the naturalness scored by MOS remains almost unchanged when less than 50% instances are pruned off, and the MOS does not severely degrade when reduction rate is above 50%.
Keywords :
speech synthesis; Corpus approach; StaRp-VPA algorithm; pruning redundant synthesis; speech synthesis; text-to-speech; Computer science; Databases; Degradation; Electronic mail; Frequency synthesizers; Hidden Markov models; Oceans; Signal processing algorithms; Signal synthesis; Speech synthesis; scalable speech synthesis system; scalable text-to-speech system; speech synthesis; text-to-speech system;
Conference_Titel :
Cybernetics and Intelligent Systems, 2008 IEEE Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-1673-8
Electronic_ISBN :
978-1-4244-1674-5
DOI :
10.1109/ICCIS.2008.4670895